The Agent Skills Directory

[COMMAND_EXECUTION]: The skill constructs and executes shell commands using the GitHub CLI (gh).
Evidence: The workflow utilizes gh issue create to submit feedback to the modu-ai/moai-cowork repository.
Risk: The command construction uses placeholders like {제목} and {본문} which are populated directly from user input. Without proper escaping by the execution environment, this could allow for argument injection within the gh command.
[INDIRECT_PROMPT_INJECTION]: The skill provides a surface for injection by processing untrusted user input and passing it to a privileged tool.
Ingestion points: User-provided bug descriptions, reproduction steps, and feature request details collected via AskUserQuestion in SKILL.md.
Boundary markers: Absent. There are no instructions for the agent to use delimiters or to treat the user input as data rather than instructions when interpolating it into the CLI command.
Capability inventory: The skill has the capability to execute the gh CLI tool to create issues on a remote repository.
Sanitization: Absent. The skill does not define any escaping, validation, or filtering logic for the content provided by the user.

meta-feedback